Hide Forgot
Created attachment 530113 [details] screenshot of the incident Description of problem: This issue is extending mouse issue from Bug 732467 as nobody was describing this issue. The cursor is not reflecting the real pointer position. Version-Release number of selected component (if applicable): spice-vdagent-0.8.1-2.el6.x86_64 spice-client-0.8.2-7.el6.x86_64 How reproducible: Steps to Reproduce: 1. start rhel62 guest with xinerama enabled (requires to set 2 qxl devices) 2. connect to it trough spice 3. do a mouse drag (just like on screenshot) Actual results: see the difference between actual position of cursor and the real position Expected results: cursor should reflect the real position Additional info: turning of vdagent fixes the issue.
Created attachment 530114 [details] xorg.conf with xinerama over two qxl's
Looking at the screenshot this seems exactly like the behavior which I would expect when using multi monitor with the agent, while not having bug 732467 fixed. Did you test this with one of the brew scratch builds which has bug 732467 fixed? If not I suggest closing this as a duplicate of bug 732467. I know that the original bug description in bug 732467 is different from what you are seeing now, but the fix for bug 732467 involves changing the way how input coordinates from absolute input devices are interpreted when you've a multiple monitor Xinerama setup, the agent with multi monitor support provides coordinates in the format matching the new agreed up on way to interpret them, so using an unfixed Xorg will result in the mouse position being off. To get the old stuck to right pointer behavior use qemu with a usb tablet and without the agent.
moving to 6.3, too late for 6.2, unfortunately